© 2019. Published by The Company of Biologists Ltd | Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300

RESEARCH ARTICLE Rho kinase-dependent apical constriction counteracts M-phase apical expansion to enable mouse neural tube closure Max B. Butler1,*, Nina E. Short1,*, Eirini Maniou1, Paula Alexandre1, Nicholas D. E. Greene1, Andrew J. Copp1 and Gabriel L. Galea1,2,‡

ABSTRACT presumptive organs. These forces may be generated non-cell- Cellular generation of mechanical forces required to close the autonomously, such as during osmotic swelling of the lumen of the presumptive spinal neural tube, the ‘posterior neuropore’ (PNP), closed neural tube, the embryonic precursor of the brain and spinal involves interkinetic nuclear migration (INM) and apical constriction. cord (Desmond and Jacobson, 1977). Morphogenetic forces are also Both processes change the apical surface area of neuroepithelial cells, cell-autonomously generated through conserved mechanisms that but how they are biomechanically integrated is unknown. Rho kinase alter the shape of cells and, collectively, tissues (Pearl et al., 2017). (Rock; herein referring to both ROCK1 and ROCK2) inhibition in mouse Probably the best studied force-generating mechanism is apical whole embryo culture progressively widens the PNP. PNP widening is constriction of epithelial cells, which requires recruitment of not caused by increased mechanical tension opposing closure, as non-muscle myosin motor proteins, such as myosin-II, onto the evidenced by diminished recoil following laser ablation. Rather, Rock apical F-actin cytoskeleton. Apical myosin recruitment is promoted inhibition diminishes neuroepithelial apical constriction, producing by the activity of Rho-associated kinase (Rock; herein referring to increased apical areas in neuroepithelial cells despite diminished both ROCK1 and ROCK2 for mammalian systems) (Das et al., tension. Neuroepithelial apices are also dynamically related to INM 2014; Mason et al., 2013; Sai et al., 2014). In Drosophila and progression, with the smallest dimensions achieved in cells positive for non-mammalian vertebrates, apical constriction proceeds in an the pan-M phase marker Rb phosphorylated at S780 (pRB-S780). asynchronous ratchet-like pulsatile manner, producing wedge- A brief (2 h) Rock inhibition selectively increases the apical area of shaped cells with narrowed apical and widened basolateral pRB-S780-positive cells, but not pre-anaphase cells positive for domains (Christodoulou and Skourides, 2015; Martin et al., phosphorylated histone 3 (pHH3+). Longer inhibition (8 h, more than 2009). When coordinated across an epithelium, this causes tissue one cell cycle) increases apical areas in pHH3+ cells, suggesting cell bending (Nishimura et al., 2012). cycle-dependent accumulation of cells with larger apical surfaces Although apical constriction has been extensively studied in during PNP widening. Consequently, arresting cell cycle progression columnar and cuboidal epithelia, its regulation and function in with hydroxyurea prevents PNP widening following Rock inhibition. highly complex pseudostratified epithelia, such as the mammalian Thus, Rock-dependent apical constriction compensates for the neuroepithelium, are comparatively understudied. Pseudostratified PNP-widening effects of INM to enable progression of closure. epithelia also undergo oscillatory nuclear migration as cells progress through the cell cycle, known as interkinetic nuclear migration This article has an associated First Person interview with the first (INM). Nuclear movement during INM is believed to proceed in authors of the paper. phases: active microtubule-dependent nuclear ascent towards the apical surface during G2 followed by actin-dependent cell rounding KEY WORDS: Rock, Posterior neuropore, Apical constriction, in M phase and ‘passive’ nuclear descent towards the basal Interkinetic nuclear migration, F-actin, Biomechanics surface during G1/S (Kosodo et al., 2011; Leung et al., 2011; Spear and Erickson, 2012). Progression of INM also influences the INTRODUCTION dimensions of the apical portion of a cell. During S phase, nuclei are Abnormalities in embryonic cellular biomechanics are increasingly basally located and the apical surface is small, mimicking apically recognised as underlying congenital structural malformations in constricted wedge-shaped cells, whereas nuclei are larger and organ systems, including the heart (Hoog et al., 2018), eye (Hosseini apically located during mitosis, presumably producing larger apical et al., 2014; Oltean et al., 2016), joints (Singh et al., 2018) and central surfaces (Guthrie et al., 1991; Nagele and Lee, 1979). nervous system (Galea et al., 2017, 2018). Mechanical forces must be Both INM and apical constriction occur in the pseudostratified generated to change the shape of embryonic structures into the neuroepithelium of the closing neural tube. Failure of neural tube closure causes severe congenital defects, such as spina bifida, in ∼ 1Developmental Biology and Cancer, UCL GOS Institute of Child Health, London 1:1000 births (Cavadino et al., 2016). Spina bifida arises due to WC1N 1EH, UK. 2Comparative Bioveterinary Sciences, Royal Veterinary College, failure of the open caudal segment of the neural tube, the posterior London NW1 0TU, UK. neuropore (PNP), to undergo the narrowing and shortening required *These authors contributed equally to this work for closure. PNP closure is fundamentally a biomechanical event ‡Author for correspondence ([email protected]) during which the flat neural plate elevates lateral neural folds that buckle at paired dorsolateral hinge points. The neural folds become G.L.G., 0000-0003-2515-1342 apposed medially, such that their tips meet at the dorsal midline This is an Open Access article distributed under the terms of the Creative Commons Attribution where they are then joined by cellular protrusions that ‘zipper’down License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed. the length of the neuropore (Nikolopoulou et al., 2017). PNP narrowing through neural fold medial apposition involves both

Received 29 January 2019; Accepted 30 May 2019 apical constriction and INM. Regional prolongation of S phase Journal of Cell Science

1 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300 in the neuroepithelium along the PNP midline results in the was significantly reduced caudal to this location (Fig. 1D). accumulation of wedge-shaped cells, bending the tissue at the Dorsolateral hinge points were still present in Rock-inhibited medial hinge point (McShane et al., 2015; Smith and Schoenwolf, embryos (Fig. 1E), as previously reported (Escuin et al., 2015). 1988). Unlike pulsatile apical constrictions, this hinge point is stable These morphometric studies reveal tissue shape changes caused by and persists at the tissue level throughout most of PNP closure 8 h of Rock inhibition, of which PNP widening is the most marked, (Shum and Copp, 1996). for which biomechanical mechanisms were further investigated. PNP closure can be expected to fail if its tissue structures are As previously reported (Escuin et al., 2015), Rock inhibition abnormal, if pro-closure cell-generated mechanical forces cannot diminished the selective localisation of F-actin in the apical exceed forces which oppose closure or if those forces are not neuroepithelium (Fig. S1). In addition, we specifically investigated transmitted in a coordinated manner across the PNP. We have recently two supracellular F-actin organisations present in the PNP (Galea reported two genetic mouse models in which excessive tissue tensions et al., 2017, 2018): long rostrocaudal cables along the neural folds opposing PNP closure predict failure of closure and development of (Fig. 2A) and mediolateral profiles identifiable in the apical spina bifida (Galea et al., 2017, 2018). Tissue tension was inferred neuroepithelium (Fig. 2C). Although rostrocaudal F-actin cables from physical incision or laser ablation experiments in which the most remained evident close to the zippering point, the proportion of the recently fused portion of the neural tube, the zippering point, was PNP not flanked by these cables was significantly greater in embryos disrupted and the resulting rapid deformation of the PNP quantified treated with 10 µM Y27632 for 8 h than those cultured in vehicle (Galea et al., 2017, 2018). These experiments also showed that the (Fig. 2A,B; also see Fig. 1A). Unexpectedly, mediolaterally oriented PNP is a biomechanically coupled structure thanks at least in part to supracellular F-actin profiles were still evident in the neuroepithelium supracellular actomyosin cables that run rostro-caudally along the tips of the open PNP of Rock-inhibited embryos (Fig. 2C). The average of the neural fold (Galea et al., 2017, 2018). Hence, ablation of the orientation of F-actin profiles in each PNP was not significantly PNP zippering point causes neuropore widening, which extends into different between vehicle and Rock-inhibited embryos (Fig. 2C,D). more posterior portions of the open region. The apical neuroepithelium also forms distinct supracellular F-actin enrichments (‘profiles’)that Rock inhibition diminishes PNP tissue tension are oriented mediolaterally, in the direction of neural fold apposition Diminished rostrocaudal F-actin cables are associated with increased (Galea et al., 2018; Nishimura et al., 2012). Consistent with the tissue tension in Zic2Ku/Ku embryos, as shown by greater tissue recoil involvement of specialised F-actin structures in PNP closure, inhibition following zippering point ablation (Galea et al., 2017). In contrast, of the actomyosin regulator Rock with the commonly used antagonist Rock inhibition for 8 h prior to and during zippering point ablation Y27632 stalls PNP closure in mice and other vertebrates (Escuin et al., substantially diminished lateral tissue recoil (Fig. 3A,B). The caudal- 2015; Kinoshita et al., 2008). open PNP of Rock-inhibited embryos appeared to narrow following Rock inhibition impairs the selective apical enrichment of zippering point ablation, although this did not reach significance at actomyosin required for apical constriction in the neuroepithelium any position. These findings suggest Rock inhibition decreases tissue (Escuin et al., 2015) and other tissues (Harding and Nechiporuk, tensions that pull the neural folds laterally. 2012; Sai et al., 2014). We set out to test whether stalling of PNP We next analysed cell-level tension by quantifying recoil from closure in Rock-inhibited embryos is caused by lack of apical targeted laser cuts of the cell borders on the neural folds along which constriction, or whether it involves failure of alternative force- the rostrocaudal cables normally run. These ablations were performed generating mechanisms, such as INM. In testing this hypothesis, close to the zippering point, where F-actin cables could typically still we investigated a more fundamental question: how are apical be identified in Rock-inhibited embryos. Ablated borders in vehicle- constriction and INM functionally coordinated to regulate apical treated embryos elongated by ∼2 µm immediately following ablation area in neuroepithelial cells? (Fig. 3C). Border recoil following laser ablation in Rock-inhibited embryos was less than half of that quantified in vehicle-treated RESULTS controls (Fig. 3C,D), corroborating reduced tension following Rock Rock inhibition widens the PNP and diminishes the neural inhibition. Collectively, these tissue- and cell-level laser ablation fold actomyosin cables experiments suggest that PNP widening in Rock-inhibited embryos is Prolonged Rock inhibition at concentrations compatible with not caused by increases in tissue tensions pulling on the zippering continued development in mouse whole embryo culture delays point. PNP closure, producing longer PNPs than in vehicle-treated embryos We next sought to determine whether Rock inhibition alters tension (Escuin et al., 2015). To minimise the potential for secondary within the neuroepithelium of the open PNP. This posed a challenge changes owing to prolonged culture, we first characterised the as, unlike simpler epithelia, neuroepithelial cells do not have morphological changes caused by 8 h of Rock inhibition with the predictable straight borders that can be reproducibly visualised and extensively used compound Y27632 in embryonic day (E)9–9.5 CD1 ablated (Fig. S2A′). We initially developed a method of performing mouse embryos. This treatment period is sufficient to observe long, linear laser ablations along the apical neuroepithelium and biologically meaningful differences in PNP dimensions (Hughes quantified lateral retraction (Fig. S2A,B). While this confirmed et al., 2018). After 8 h of Rock inhibition, we observed dose- the neuroepithelium is under tension, recoil magnitudes varied dependent widening of the PNP, giving rise to PNPs that were more substantially along the length of the ablation (Fig. S2B). A pilot ‘diamond-shaped’ as opposed to the the elliptical structures study of five comparable embryos ablated on the same day showed this characteristic of control embryos at late stages of closure method was too variable to allow meaningful comparisons between (Fig. 1A,B). This short period of Rock inhibition did not treatment groups (sample size calculation based on pilot study significantly increase PNP length (Fig. 1C). Neural fold elevation quantifying mid-ablation recoil in five embryos required 90 embryos tended to be more variable in Rock-inhibited than vehicle-treated to detect a 30% difference with 80% power at P<0.05). To circumvent embryos (Leven’stestP=0.10), but was not significantly altered by these issues, we developed a novel method in which a 30-µm-diameter Rock inhibition in the region near the zippering point (25% of the annular ablation is created in the neuroepithelium, isolating a cluster of length of the PNP from the rostral end, Fig. 1D). Neural fold elevation approximately eight cells (Fig. 4A). The reduction in area of this Journal of Cell Science

2 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300

Fig. 1. Rock inhibition widens the PNP and reduces neural fold elevation. E9 CD1 embryos were cultured in vehicle (Veh, n=6) or with the indicated concentrations of Y27632 (5 µM, n=5; 10 µM, n=6) for 8 h. (A) Representative wholemount phalloidin-stained vehicle and 10 µM Y27632-treated embryo PNPs (dorsal view). Scale bar: 100 µm. Image is presented after applying an inverted greyscale look-up table. (B) Sequential quantification of PNP width at every 1% of its length. The schematic in B′ illustrates sequential width measurements shown by the cyan lines across the PNP. (C) Quantification of PNP length as shown by the cyan line in the schematic. (D) PNP elevation was quantified as the dorsoventral distance between the neural fold tips and apical surface of the midline neuroepithelium (vertical line in E) at 25%, 50% and 75% of the length of the PNP. (E) 3D-reconstructed images of a vehicle and 10 µM Y27632-treated embryo PNP, illustrating the presence of dorso-lateral hinge points in both. These reconstructions are shown looking rostrally into the closure neural tube as indicated by the cyan arrow in the schematic. The red asterisks denote the zippering point throughout. **P<0.01 (tests defined in Materials and Methods); embryos were analysed at the 19–22 somite stage.

cluster of cells as they constrict immediately following ablation was is not caused by ‘stretching’ given tissue tension is diminished. The then quantified as a readout of tension (Fig. 4A–C, sample size simultaneous increase in apical size and reduction in mechanical calculation based on pilot study of five embryos required tension is consistent with Rock inhibition stopping apical approximately eight embryos to detect a 30% difference with 80% constriction in at least a subpopulation of cells in the mammalian power at P<0.05). Cell constriction following annular ablations was neuroepithelium. significantly greater in vehicle-treated than Rock-inhibited embryos, suggesting that Rock inhibition reduces tension in the neuroepithelium The apical area of neuroepithelial cells decrease during exit (Fig. 4C). from M phase To establish whether this reduction in neuroepithelial tension The presence of larger apical surfaces in neuroepithelial cells, which correlates with diminished apical constriction, apical areas were take up more space, may explain why Rock inhibition widens the quantified in N-cadherin-stained PNPs (Fig. 4D). The median apical PNP. Given that apical surfaces widen as nuclei approach the apical area for the neuroepithelial cells were significantly larger in Rock- side during INM, we hypothesised this may be related to inhibited than in vehicle-treated embryos (Fig. 4E). However, the accumulation of mitotic cells with large apical surfaces in Rock- distribution of apical area observed in this epithelium was highly inhibited embryos. Rock1 protein is normally enriched around the skewed. Frequency versus apical areas plots were used to analyse apices of neuroepithelial cells, including mitotic cells positive for shifts in apical area across the cell population. Leftward shifts in phosphorylated histone 3 (pHH3) (Fig. S3A), but Rock inhibition the frequency curve indicate a greater proportion of cells in the did not substantially alter the neuroepithelial mitotic index population have small apical areas, whereas rightward shifts indicate (Fig. S3B). Furthermore, neuroepithelial cells with the largest a greater proportion of cells have large apical areas. In Rock- observed apical areas were not necessarily positive for the G2/M inhibited embryos the proportion of cells with large apical areas phase marker pHH3 (Fig. 5A). This was visualised in uncultured increased, producing a significant shift in observed dimensions embryos triple-labelled in whole-mount for pHH3, ZO-1 (also towards larger sizes (short arrow in Fig. 4F), although the majority known as TJP1) to show the apical surface, and scribble (Scrib) of cells retained small apical areas despite Rock inhibition (long to label basolateral cell borders so that individual nuclei could arrow in Fig. 4F). Taken together, these data suggest that the be definitively related to their own apical surface. Scrib increase in the neuroepithelial apical size in Rock-inhibited embryos recapitulated ZO-1 staining at the apical surface (Fig. 5A), so Journal of Cell Science

3 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300

Fig. 3. Rock inhibition diminishes anti-closure PNP tissue tension. (A) Representative reflection live-imaged PNPs following 8 h of culture in Fig. 2. Rock inhibition diminishes the rostrocaudal neural fold F-actin vehicle or 10 µM Y27632. Each neuropore was imaged before and again after cables but not mediolateral neuroepithelial profiles. E9 CD1 embryos laser ablation (red line) of the zippering point. The white perimeter indicates the were cultured in vehicle or the indicated concentrations of Y27632 for 8 h. shape of the PNP before ablation, the green-shaded region indicates lateral (A) Representative wholemount-stained vehicle- and 10 µM Y27632-treated displacement of the neural folds. Ablations of the rostrocaudal F-actin cables embryo PNPs (dorsal view). Arrowheads indicate the rostrocaudal were performed in different embryos in the region indicated by A′. Scale bar: actomyosin cables. (B) Quantification of the proportion of the PNP which 100 µm. The red asterisks denote the zippering point. (B) Mediolateral change extends beyond the caudal limit of the rostrocaudal cables, as previously in width of the PNP in vehicle- (n=7) and 10 µM Y27632-treated embryos (n=6) defined (Galea et al., 2017; Hughes et al., 2018). Vehicle, n=8; 5 µM, n=5; at each 1% of the length of the PNPs from the zippering point (position 0%). 10 µM, n=7. (C) Visualisation of mediolateral F-actin profiles (image is Green lines indicate the region in which vehicle-treated embryos recoiled to a presented after applying an inverted grey look-up table, which is binarised in significantly greater extent than 10 µM Y27632-treated embryos (P<0.05, the magnified views) in the relatively flat portion of the PNP in vehicle- and mixed model testing). (C) Rostrocaudal change in length (example in C′) of cell 10 µM Y27632-treated embryos. Dashed lines indicate the mean orientation borders along the neural folds following laser ablation. n=9 per group, **P<0.01 quantified for those embryos. (D) Quantification of the mean orientation of (t-test). (D) Representative kymographs (from cell borders equivalent to the binarised F-actin profiles in vehicle (n=8) and 10 µM Y27632-treated cyan box in C′) of cable ablations in vehicle- and 10 µM Y27632-treated embryos (n=8) relative to the mediolateral direction (cyan angle bracket in the embryos. Bright spots in each temporal slice are cell membranes on either side schematic). The red asterisks denote the zippering point throughout. Scale (left and right) of the ablation. White arrows indicate the ablated border, and bars: 100 µm. **P<0.01 (ANOVA with post-hoc Bonferroni); embryos were temporal slices below the horizontal cyan line show displacement after laser analysed at the 19–22 somite stage. ablation. Scale bar: 5 µm, kymographs were generated in Fiji. Journal of Cell Science

4 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300

Fig. 4. Rock inhibition reduces neuroepithelial apical constriction. (A) Representative annular ablation in the region indicated by the white circle in the representative whole-PNP view, imaged before and immediately after ablation. A single ablation was performed in each embryo. The area within the ablated circle is shown by the cyan polygon between definable landmarks before ablation, which deformed to the magenta polygon immediately following ablation. Scale bar: 25 µm. (B) Representative segmented and registered cell borders before (cyan) and immediately after (magenta) annular ablation (white circle) illustrating their displacement in both vehicle- and 10 µM Y27632- treated embryos. The red arrows illustrate that the surrounding tissue also retracts away from the ablation. (C) Quantification of the constriction of the tissue within the ablated circle in vehicle- (n=8) and 10 µM Y27632-treated (n=9) embryo PNPs. (D) Surface- subtracted N-cadherin staining from a vehicle-treated embryo. Scale bar: 20 µm. Cell borders were segmented using Tissue Analyser as illustrated. (E) Quantification of median apical areas of neuroepithelial cells based on segmented N-cadherin staining in vehicle- and 10 µM Y27632-treated embryos (n=6 each) following 8 h of culture. (F) Frequency plot of observed apical areas of neuroepithelial cells in vehicle- (324 cells from 6 embryos) and 10 µM Y27632-treated embryos (244 cells from six embryos). The arrows indicate that although the majority of cells retain small apical areas despite Rock inhibition, there is a highly significant shift towards more cells having large apical areas. *P<0.05, **P<0.01, **P<0.001 (tests defined in Materials and Methods).

Scrib staining was used to analyse apical area in subsequent cells (Fig. 5C, mean of 20.4 µm2). A small proportion of pS780+ cells experiments. These analyses showed that although pHH3 will have recently completed division. sometimes labels cells with a large apical area, it can also label These findings suggest that apical areas of neuroepithelial cells cells with small apices (Fig. 5A). 3D cell reconstructions showed typically decrease as cells transition through M phase towards G1. that pHH3+ cells can have small apical areas relative to their pHH3- In order to dynamically visualise this apical re-constriction at the negative neighbours. Non-mitotic neighbours appear to wrap over end of mitosis, we live-imaged the hindbrain neuroepithelium of pHH3+ cell bodies (Movie 1), maintaining continuity of apical zebrafish embryos (Fig. 5D). Mitotic cells were identified as those junctions. which completed division to form two daughter cells while being The commonly used S10-phosphorylated pHH3 is not a pan-M imaged and their apical areas over ∼20 min prior to division were phase marker; cells are positive for this marker from G2 to anaphase analysed (Fig. 5D). In this species, we found that apical sizes in (Dai et al., 2005; Ren et al., 2018). To extend our analysis we stained neuroepithelial cell increased to a maximum size in early mitosis, for the pan-M phase phosphorylated S780 epitope of retinoblastoma then rapidly re-constricted prior to division (Fig. 5E). Thus, end- protein (an Aurora B kinase target site; Macdonald and Dick, 2012; mitotic apical re-constriction is an evolutionarily conserved Nair et al., 2009), which produces bright staining (hereafter denoted neuroepithelial cell behaviour in fish and mice. pS780+) throughout M phase, persisting beyond anaphase to the end of cytokinesis (Jacobberger et al., 2008) (see Fig. 5B). In order Rock inhibition preferentially increases apical area in late to more closely analyse the distribution of apical areas around the M phase cells time of mitosis, we triple-labelled PNPs for Scrib, pHH3 and To determine whether Rock-dependent apical constriction is pS780. specific to a cell cycle phase, we initially analysed the apical The apical areas of pHH3+ (G2 to anaphase) or pS780+ (M phase area of pHH3+ or pS780+ cells after 2 h of Rock inhibition. This to cytokinesis) neuroepithelial cells were analysed in the relatively flat duration of treatment was selected because it is sufficient for region of the PNP caudal to the median hinge point of non-cultured significant PNP narrowing to occur (Galea et al., 2017), but not for mouse embryos. Single versus double positivity for these markers cells to progress through a complete cell cycle (McShane et al., could not be taken into account given the small number of cells 2015). This Rock inhibition was sufficient to increase the labelled with either, typically ∼15 per PNP. pHH3+ cells included proportion of cells with large apical areas in the overall those with the largest apical area observed, but the majority had small population (Fig. 6A, All). Remarkably, the apical areas of apical surfaces (Fig. 5C, mean apical area for ‘All’ cells, 34.3 µm2; pHH3+ cells were not significantly altered by Rock inhibition pHH3+,26.0µm2). The distribution of pS780+ apical surfaces was (Fig. 6A, pHH3+), whereas apical areas of pS780+ cells were + significantly shifted towards smaller dimensions relative to pHH3 significantly larger in Rock-inhibited than vehicle-treated Journal of Cell Science

5 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300

Fig. 6. Increases in apical area of neuroepithelial cells following Rock inhibition are cell cycle stage specific. (A,B) Embryos were cultured for (A) 2 h or (B) 8 h in vehicle or 10 µM Y27632 and triple-stained for Scrib, Fig. 5. Neuroepithelial cells undergo apical re-constriction in late M phase. pHH3 and pS780. Apical areas were analysed in the overall cell population A–C represent data from non-cultured mouse PNPs, whereas D and E are from (‘All’), in all pHH3+ cells and in all pS780+ cells. B′ illustrates the shape live-imaged zebrafish hindbrain neuroepithelium. (A) Representative triple- (red, Scrib-labelled cell borders) and apical area (white rings) of a pHH3/pS780 labelled neuroepithelial apical surface showing ZO-1 (apical-most tight junction double-positive cell from a vehicle- and Y27632-treated embryo. n numbers marker), Scrib and the G2/M phase marker pHH3. Circles indicate cells with large are as follows: All, 2 h vehicle, n=310 cells from six embryos, Y27632, n=267 apical areas that are negative for pHH3. Squares indicate pHH3+ cells with large cells from six embryos; All, 8 h vehicle, n=253 cells from five embryos, Y27632, and small apical areas, as shown in the magnified views. (B) Representative n=320 cells from seven embryos; pHH3+, 2 h vehicle, n=87 cells, Y27632, wholemount maximum projection showing the pattern of Scrib, pHH3 and pS780 n=80 cells; pHH3+, 8 h vehicle, n=110 cells, Y27632, n=102 cells; pS780+,2h staining in an uncultured embryo. The dashed white box indicates the relatively vehicle, n=118 cells, Y27632, n=97 cells; pS780, 8 h vehicle, n=68 cells, flat region of the PNP caudal to the medial hinge point in which apical areas were Y27632, n=84 cells. (C) Schematic of the proposed model of apical analysed. Optical cross-sections through pHH3/pS780 single and double constriction of neuroepithelial cells as cells transition from early M (pHH3+/ positive cells (which appear cyan) are also shown. The white asterisk denotes the pS780−) through cytokinesis into G1 (pHH3−/pS780+). Rock-dependent zippering point. Scale bar: 100 μm. (C) Frequency plot showing the distribution of constriction is indicated in late M phase. (D) Sequential quantification of PNP apical area (based on Scrib staining) in the overall neuroepithelial cell population width at every 1% of its length in embryos cultured in vehicle (2 h culture) or (‘All’, 262 cells from 5 embryos), and for pS780-positive (69 cells) and pHH3- 10 µM Y27632 for 2 h or 4 h (n=6 embryos per group analysed at the 16–20 positive cells (58 cells). **P<0.01 (Kolmogorov–Smirnov test). (D) Representative somites stage). *P<0.05, **P<0.01, ***P<0.001 (tests defined in Materials and snapshots of a live-imaged zebrafish hindbrain neuroepithelium, with the apical Methods). cell surface mosaically labelled with Par3–RFP (cyan outline), undergoing apical reconstriction prior to division. The last time point shown (post-division) was not included in the apical area analyses shown in E. (E) Quantification of Following 8 h of treatment, Rock inhibition significantly increased apical area for cells in the zebrafish hindbrain neuroepithelial over time. apical areas overall as well as in both the pHH3+ and pS780+ For each cell, its maximum size prior to division was identified (set at 100% for populations (Fig. 6B). Rock inhibition did not significantly change each cell) and 10 time points were analysed around this maximum dimension (i.e. t=10 min set at 100%). n=14 divisions from seven embryos. ***P<0.001 the proportion of cells labelled with pHH3 or pS780 individually versus the maximum dimension (repeated measures ANOVA with Bonferroni or together (Fig. S3C), further suggesting that Rock inhibition post-hoc). minimally affects mitotic progression in this epithelium. Blocking the Rock-dependent reduction in apical area as cells transition from G2 to embryos (Fig. 6A, pS780+). Consequently, following Rock G1 may therefore have cumulative effects as cells continue to inhibition the observed apical area frequencies of pS780+ cells progress through the cell cycle (Fig. 6C), progressively widening the approximated those of pHH3+ cells (Fig. 6A, pHH3+ versus PNP. Consistent with this, whereas 2 h of Rock inhibition was + pS780 ). sufficient to significantly widen the PNP (or prevent its narrowing in Journal of Cell Science

6 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300 culture), PNP width increased further following an additional 2 h of inhibition (4 h total, Fig. 6D).

Blocking cell cycle progression prevents PNP widening caused by Rock inhibition The findings that Rock inhibition increases neuroepithelial apical areas at specific cell cycle phases in a temporally restricted manner suggest that progression through the cell cycle contributes to PNP widening in Rock-inhibited embryos. This predicts that inhibition of progression through the cell cycle may diminish PNP widening following Rock inhibition. To test this, we treated embryos with the ribonucleotide reductase inhibitor hydroxyurea (HU) which blocks cell entry into S phase (Leitch et al., 2016; Philips et al., 1968). As well as being used therapeutically for various conditions in humans, HU has previously been reported to reduce spina bifida incidence in the curly tail mouse model in vivo (Seller and Perkins, 1983). In the present study, treatment with 0.8 mM HU for 8 h substantially, but not completely, diminished the neuroepithelial mitotic index (Fig. S4A,B). Reducing cell cycle progression with HU did not restore the rostrocaudal F-actin cables (Fig. 7B), but fully prevented PNP widening after 8 h of Rock inhibition (Fig. 7A,C). This is consistent with a requirement for cell cycle progression during PNP widening due to Rock inhibition. Diminished cell cycle progression in HU-treated embryos precluded analysis of cell cycle phase-specific apical areas. Given that HU treatment blocks cells in S phase (Leitch et al., 2016; Philips et al., 1968), when they have small apical areas (Guthrie et al., 1991; Nagele and Lee, 1979), a greater proportion of neuroepithelial cells had small apical areas after 8 h of HU treatment (Fig. 7D), as expected. Rock inhibition increased apical areas in a subset of neuroepithelial cells independantly of cell cycle progression, but HU treatment partly prevented this increase (Fig. 7E). Thus, blocking cell cycle progression produces fewer cells (fewer divisions) with smaller apical sizes than Rock inhibition alone, preventing progressive PNP widening. Fig. 7. Cell cycle progression is a prerequisite for PNP widening in Rock- inhibited embryos. (A) Representative 3D-rendered PNP images from DISCUSSION embryos treated with vehicle, 10 µM Y27632 (Y2), 0.8 mM HU or HU+Y27632 Tissue-level integration of cellular force-generating mechanisms is after 8 h of culture. The asterisks indicate the zippering point, cyan lines necessary to achieve coordinated morphogenetic shape change. approximate the mid-PNP width. The insets below the HU and HU+Y27632- Force-generating mechanisms often act in opposing ways; for treated embryos are presented after applying an inverted grey look-up table for example, both cell apoptosis (Oltean and Taber, 2018) and regional phalloidin staining to facilitate visualisation of the rostrocaudal F-actin cables cell proliferation (Hosseini et al., 2017; Peeters et al., 1998) can (arrows). Scale bars: 100 µm. (B) Quantification of the proportion of the PNP change tissue shape despite having opposing effects on cell number. which extends beyond the rostrocaudal cables (as in Fig. 2B) in each treatment group. (C) Quantification of mid-PNP width in each treatment group. Embryos This mutual antagonism may also be true of the interplay between were analysed at the 20–23 somites stages; vehicle, n=6; Y27632, n=7; HU, apical constriction and INM. Here, we assessed the functional n=9; HU+Y27632, n=9. (D,E) Apical areas of neuroepithelial cells were integration of these force-generating mechanisms in the mammalian analysed in ZO-1-stained PNPs from (D) vehicle versus HU-treated and neuroepithelium by using a highly reproducible model of PNP (E) 10 µM Y27632- versus HU+Y27632-treated embryos after 8 h of culture. closure suppression through pharmacological inhibition of Rock. n numbers were: vehicle, n=845 cells from six embryos, HU, n=816 cells from Pharmacological antagonism allows greater temporal control over seven embryos; Y2, n=861 cells from six embryos, HU+Y2, n=901 cells from six embryos. NS, not significant; *P<0.05, **P<0.01, ***P<0.001 (tests defined Rock activity than can currently be achieved genetically in mouse in Materials and Methods). embryos. Small molecule Rock inhibitors are in clinical development for conditions ranging from glaucoma to cardiovascular disease (Hartmann et al., 2015; Honjo and Tanihara, 2018). In the present mediolaterally oriented F-actin profiles remained evident in study, we demonstrate that Rock is required for supracellular the neuroepithelium of Rock-inhibited mouse embryos. This is organisation of F-actin into biomechanically coupling rostrocaudal consistent with the recent finding that directional supracellular cables, and document marked progressive PNP widening following F-actin enrichments in stretched Drosophila wing disks also form Rock inhibition. independently of Rock (Duda et al., 2019). PNP mediolateral Rock inhibition diminishes F-actin radial and stress fibres in profiles require planar cell polarity (PCP) signalling (Galea et al., immature epidermis-derived epithelia, but has minimal effects on 2018; McGreevy et al., 2015) and, while Rock has been suggested F-actin organisation in epithelia matured in vitro, suggesting Rock- to mediate downstream events of PCP signalling in some contexts independent F-actin organisation in epithelial cells with mature (Winter et al., 2001), our findings suggest that this PCP-regulated intercellular junctions (Vaezi et al., 2002). In the present study, event is relatively independent of Rock signalling. Conditional Journal of Cell Science

7 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300 mosaic deletion of the mammalian core PCP component Vangl2 Rock-inhibited neuroepithelial cells in the present study. Our (Galea et al., 2018), or compound heterozygous mutations of findings confirm that mitotic neuroepithelial cells include those Vangl2 and the Diaphanous-related formin Daam1, lead to spina with the largest observed apical area in this epithelium. bifida (Lopez-Escobar et al., 2018). Diaphanous 1 is required for Unexpectedly, however, the mitotic population also includes Rock-independent polarised F-actin profiles to form in stretched neuroepithelial cells with some of the smallest apical areas, Drosophila wing disks (Duda et al., 2019), suggesting a Rock- leading us to suggest that an apical re-constriction event happens independent PCP pathway directs mechanoresponsive actomyosin during mitosis. Live-imaging of the zebrafish hindbrain confirmed organisation. this mitotic apical re-constriction occurs in neuroepithelial cells, and PCP mutations prevent convergent extension movements, but cell suggests it is conserved between two vertebrate species and migration is unlikely to substantially contribute to the rapid increase anatomical sites. in PNP width in Rock-inhibited embryos. PNP widening could be Mitotic cell identification using pHH3 staining is limiting caused by increases in laterally tethering mechanical tensions, as because this marker does not persist throughout M phase (Dai inferred from rapid lateral recoil of the neural folds following laser et al., 2005). Bright staining for the pS780 epitope used here ablation of the zippering point. Increased recoil precedes failure of overlaps with pHH3 in M (but not G2) phase and persists to the end PNP closure and development of spina bifida in Zic2Ku/Ku embryos of cytokinesis (Jacobberger et al., 2008). pS780+ cells have highly and in embryos with a conditional knockout in Vangl2 (Galea et al., constricted apical areas, which occurs though a Rock-dependent 2017, 2018). The opposite is seen in Rock-inhibited embryos; recoil mechanism, such that 2 h of Rock inhibition reverts the distribution is substantially diminished compared with vehicle-treated controls. of their apical areas to reflect that of pHH3+ cells. Remarkably, These experiments may be limited by differences in tissue material apical areas of pHH3+ neuroepithelial cells are insensitive to short- properties, potentially including a reduction in neuroepithelial term Rock inhibition, suggesting that the Rock-dependent apical material stiffness in Rock-inhibited embryos (Nagasaka et al., constriction process occurs at the end of M-phase. Rock- 2016). However, structural stiffness of another embryonic structure, independent apical constriction has previously been described the Xenopus blastopore, is insensitive to short-term Rock inhibition during Drosophila ventral furrow formation: pulsed treatment with (Feroze et al., 2015). Structural differences in PNP morphology Y27632 stopped the coordinated apical constriction required for might also confound tissue-level analysis, although Rock inhibition furrow formation, but permitted the myosin II-independent minimally altered structural features, such as dorsolateral hinge slow reductions in apical area concomitant with apicobasal points and neural fold elevation close to the zippering point. Our nuclear migration (Krajcovic and Minden, 2012). Possible Rock- cell-level analyses also demonstrated reduced recoil following laser independent mechanisms by which pHH3+ cells achieve small ablation of the cell borders along which the rostrocaudal actomyosin apical areas also include transfer of elastic energy (generated cables run. Taken together, these findings exclude an increase in during G2 nuclear ascent) from neighbouring cells (Shinoda et al., tissue tensions as a biomechanical explanation for PNP widening in 2018), consistent with the observed ‘wrapping’ of non-mitotic Rock-inhibited embryos. neighbours over pHH3+ cells. Alternatively, apical area may As well as reducing laterally tethering tension, Rock inhibition decrease in late M phase due to division of apical end feet into also diminished apical constriction of neuroepithelial cells. Apical two daughter cells during cytokinesis. The absence of binucleated constriction of epithelial cells is commonly documented by cells in this and previous (Escuin et al., 2015) studies suggests that visualising reductions in apical areas, as we and others have Rock inhibition did not cause failure of cytokinesis in the previously reported in the neuroepithelium (Bush et al., 1990; Galea neuroepithelium. et al., 2017; McGreevy et al., 2015). This cannot differentiate active The apical area of pHH3+ cells did increase significantly force-generating apical constriction from external compression, or following 8 h of Rock inhibition. Thus, in addition to the acute indeed cell shape changes linked to cell cycle progression in a loss of apical constriction in late M phase, neuroepithelial cells with pseudostratified epithelium. Here, we provide three levels of large apical areas accumulate as cells progress through the cell cycle evidence showing that the cells from mammalian neuroepithelium while Rock activity is inhibited. The requirement for cell cycle undergoes apical constriction in a Rock-dependent manner: (1) progression to drive PNP widening in Rock-inhibited embryos is apical F-actin localisation is lost, and (2) neuroepithelial apical confirmed by the striking rescue achieved with HU treatment. areas are larger despite (3) diminished neuroepithelial tension in Achieving rescue of PNP widening caused by Rock inhibition, Rock-inhibited embryos. Previous work from our group has shown which is both ubiquitous and pleiotropic, demonstrates the potential that inhibition of ATPase-dependent myosin-II activity with for a unified biomechanical understanding of morphogenesis to Blebbistatin rescues neuroepithelial F-actin apical localisation in identify preventative interventions for structural malformations Rock-inhibited embryos (Escuin et al., 2015). Blebbistatin itself did including neural tube defects. Taken together, our findings suggest a not impair PNP shortening, suggesting that sub-apical redistribution model of PNP neural fold apposition in which INM normally tends of F-actin potentially diminishes PNP closure by rendering the to widen the PNP due to the presence of large apical nuclei in neural plate stiff and resistant to morphogenesis (Escuin et al., G2 and early M phase. INM is counteracted by Rock-dependent 2015). In the current study, we found that Rock inhibition globally apical constrictions as cells exit M phase, maintaining apical and acutely diminishes PNP tension. This is consistent with the neuroepithelial tension. importance of actomyosin in establishing tissue tension prior to substantial extracellular matrix assembly in mammalian embryos, as MATERIALS AND METHODS previously reported in lower vertebrates (Porazinski et al., 2015). It Embryo culture and treatments remains unknown whether tissue tension in turn influences cell Studies were performed under the regulation of the UK Animals (Scientific differentiation, as it does in other contexts (Martino et al., 2018). Procedures) Act 1986 and the Medical Research Council’s Responsibility in Rock activity also regulates cell proliferation and cytokinesis in the Use of Animals for Medical Research (1993). Outbred CD1 mice were various contexts (Hazawa et al., 2018; Liang et al., 2019; Short bred in-house. Mice were mated during the day, and noon of the day a plug et al., 2017), but no difference in mitotic indices were observed in was found was considered E0. Pregnant females were killed in the morning Journal of Cell Science

8 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300 of E9 (∼16 somites at the start of culture) and their embryos were cultured Laser ablation for 2–8 h. Embryo culture was performed using the roller bottle culture Zippering point laser ablations were performed as previously described system in neat rat serum essentially as previously described by our group using a MaiTai laser (SpectraPhysics Mai Tai eHP DeepSee multiphoton (Copp et al., 2000; Hughes et al., 2018). Embryos from each litter were laser, 800 nm wavelength, 100% laser power, 65.94 µs pixel dwell time, approximately size-matched into groups, which were then randomly 1 iteration). Reflection images of live embryo PNPs were obtained using a allocated to treatment groups using coin flips. Pharmacological agents 10×/NA 0.5 Plan Apochromat dipping objective (633 nm laser wavelength). were thoroughly mixed in culture rat serum prior to adding embryos. At the PNPs were imaged before and immediately after ablation, taking ∼3 min to end of culture, embryos were dissected out of their extraembryonic capture each z-stack. membranes in the rat serum they were culture in, rinsed in ice-cold PBS and Whereas tissue-level zippering point ablations are intended to fixed in 4% PFA. compromise a relatively large region of tissue quickly (∼250 µm long, Zebrafish wild-type (AB/Tübingen) embryos were raised at 28.5°C in fish ∼30 µm deep), cable and annular ablations were optimised to ensure water or E2 medium containing 0.003% 1-phenyl-3-(2-thiazolyl)-2- targeted ablation of cell borders without vaporisation. Cable ablations were thiourea (Sigma). performed along a straight line of 0.1 µm wide at 710 nm wavelength, 80% Y27632 was purchased from Cambridge Biosciences (SM02-1) and laser power and 0.34 µs pixel dwell time for 20 iterations. Annular ablations hydroxyurea was purchased from Sigma-Aldrich (H8627-1G). Both were were performed along a 30 µm diameter ring at 710 nm wavelength, 80% dissolved in Milli-Q (MQ) water (vehicle). The concentrations and duration laser power and 0.34 µs pixel dwell time for 10 iterations. Vehicle- and of treatment with each compound is stated in the results or figure legends. Y27632-treated embryos in each experiment were alternately ablated in each experiment. Embryos were positioned in wells cut into agarose submerged Wholemount staining, confocal microscopy and image analysis in DMEM with 10% FBS immediately prior to ablation. Treated embryos Embryo wholemount staining and imaging were as previously described were kept in 10 µM Y27632 throughout, including during the ablation. (Galea et al., 2017). Alexa-Fluor-568-conjugated Phalloidin was from Thermo Fisher Scientific (A12380), rabbit anti-MHC-IIb was from Statistical analysis BioLegend (909901), goat anti-Scrib was from Santa Cruz Biosciences Comparisons between two groups were by undertaken with a Student’s (SC-11049), rabbit anti-pS780-pRB (ab47763) and rabbit anti-Rock1 unpaired t-test accounting for homogeneity of variance in Excel or in SPSS (ab45171) were from Abcam, rabbit anti-ZO-1 was from Thermo Fisher (IBM Statistics 22). Comparison of multiple groups was undertaken with a ‘ ’ Scientific (402200), mouse anti-pS10-HH3 ( pHH3 , 9706S), and mouse one-way ANOVA or Kruskal–Wallis with post-hoc Bonferroni in OriginPro anti-N-cadherin (14215S) were from Cell Signalling Technology, all as 2016 (Origin Labs). Multivariate analysis for serial PNP width or change in previously validated by the manufacturers. All primary antibodies were used width measurements (following zippering point ablation) were undertaken – at 1:100 1:200 dilution. For N-cadherin and Rock1 staining, antigen with the linear mixed models in SPSS, accounting for the fixed effects of retrieval was first performed by heating for 60 min on a 100°C hot plate in treatment and percentage of PNP length in repeated measures from each, 10 mM sodium citrate with 0.05% Tween 20, pH 6.0. Alexa Fluor- with a post-hoc Bonferroni, as previously described (Galea et al., 2017). conjugated secondary antibodies were from Thermo Fisher Scientific. Frequency distributions are plotted in 20 µm bins and were compared using Images were captured on a Zeiss Examiner LSM880 confocal using a Kolmogorov–Smirnov tests. All images are representative of embryos from 20×/NA 1.0 Plan Apochromat dipping objective. Whole PNP images were at least three independent experiments (defined as different litters processed typically captured with x/y pixel sizes of 0.59 µm and a z-step of 1.0 µm on different days). Graphs were made in OriginPro 2016 (Origin Labs) and (speed, 8; bidirectional imaging, 1024×1024 pixels). Images to analyse are represented as box plots, or as the mean±s.e.m. when several groups are apical areas were captured with x/y pixel sizes of 0.21 µm and a z-step of shown per measurement level. For box plots, the box represents the 25–75th 0.68 µm. Images were processed with Zen2.3 software and visualised as percentiles, and the median is indicated by a line and the mean by a square maximum projections in Fiji or as 3D reconstructions in Icy (whole PNPs) symbol. The whiskers show the 95% confidence intervals, and outliers are or Mesh Lab (individual cells) software. indicated. P<0.05 was considered statistically significant. To quantify the apical area of neuroepithelial cells in N-cadherin-stained PNP wholemounts, z-stacks were first surface subtracted to only show the Acknowledgements apical 2–3 µm of tissue, and cell borders were segmented using Tissue The authors wish to thank Dr Matteo Mole for critical discussions as well as Dr Lucy Analyser (Aigouy et al., 2016) as previously described (Galea et al., 2018; Culshaw and Rosie Marshall for technical assistance. macro available at https://www.ucl.ac.uk/child-health/core-scientific- facilities-centres/confocal-microscopy/publications). N-cadherin staining Competing interests does not clearly demarcate cell borders, which is necessary to relate the A.J.C. acts as a paid consultant for ViiV Healthcare Limited, with fees going to apical area to cells in specific cell cycle stages. To do this, Scrib was used support his research programme. No other competing interests are declared. and apical areas were analysed in full z-stacks by identifying and manually drawing around individual apical surfaces of cells positive for either Author contributions pHH3 or pS780. The marked differences in PNP morphology observed in Conceptualization: M.B.B., N.D.E.G., A.J.C., G.L.G.; Methodology: P.A., G.L.G.; Rock-inhibited embryos negated blinding to treatment group. Validation: N.E.S., G.L.G.; Formal analysis: M.B.B., N.E.S., E.M., G.L.G.; Investigation: M.B.B., N.E.S., E.M., P.A., G.L.G.; Resources: N.D.E.G., A.J.C., Morphometric comparisons were made using standard length measuring G.L.G.; Data curation: M.B.B., N.E.S., G.L.G.; Writing - original draft: M.B.B., N.E.S., tools in Fiji. To analyse mediolateral F-actin profile enrichment, phalloidin- N.D.E.G., A.J.C., G.L.G.; Writing - review & editing: P.A., E.M., N.D.E.G., A.J.C., stained neuroepithelial wholemount maximum projections were first G.L.G.; Visualization: M.B.B., N.E.S., P.A., G.L.G.; Supervision: G.L.G.; Project segmented by performing local contrast enhancement (CLAHE: 127 administration: N.D.E.G., A.J.C., G.L.G.; Funding acquisition: P.A., N.D.E.G., blocksize, 256 histogram bins on 16-bit images, 3 maximum slope), and A.J.C., G.L.G. then were binarised and despeckled. Average profile orientation was then calculated using the fit ellipse function in Fiji. Funding To live-image neuroepithelial divisions, zebrafish embryos were injected This study was funded partly by a Wellcome Trust Postdoctoral Clinical Research with Par3–GFP and Par3–RFP mRNAs as explained in Alexandre et al. Training Fellowship (107474/Z/15/Z) and partly by a Wellcome Clinical Research (2010). Embryos at 24–30 h post fertilisation (hpf) were anaesthetised in Career Development Fellowship (211112/Z/18/Z), both to G.L.G. N.E.S. was funded by a University College London (UCL) Child Health Research studentship and MS-222 (Sigma), immobilised in 1% low-melting-point agarose and M.B.B. by a Wellcome Biomedical Vacation Studentship. A.J.C. and N.D.E.G. imaged using a LSM 880 (Zeiss) laser scanning confocal microscope and acknowledge funding from the Wellcome Trust (087525 to A.J.C., N.D.E.G.), the 20×/NA 0.95 water immersion objective. A series of small z-stacks (planes Medical Research Council (J003794 and K022741 to N.D.E.G. and A.J.C.) and the between 0.2–1 μm apart) were obtained every 1 to 2.5 min for 1–2h. Bo Hjelt Spina Bifida Foundation (to A.J.C.). A.J.C. and N.D.E.G. are supported by Data sets were prepared using Huygens deconvolution software and Great Ormond Street Hospital Charity. P.A. was funded by the Royal Society neuroepithelial apical areas were analysed by using Fiji software. (DH100213). This research was also supported by the National Institute for Health Journal of Cell Science

9 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300

Research (NIHR) Great Ormond Street Hospital Biomedical Research Centre. The ROCK-dependent phosphorylation of NUP62 regulates p63 nuclear transport and views expressed are those of the author(s) and not necessarily those of the NHS, the squamous cell carcinoma proliferation. EMBO Rep. 19, 73-88. doi:10.15252/ NIHR or the Department of Health. Deposited in PMC for immediate release. embr.201744523 Honjo, M. and Tanihara, H. (2018). Impact of the clinical use of ROCK inhibitor on the pathogenesis and treatment of glaucoma. Jpn. J. Ophthalmol. 62, 109-126. Data availability doi:10.1007/s10384-018-0566-9 Raw values used to generate graphs shown in this manuscript are available at http:// Hoog, T. G., Fredrickson, S. J., Hsu, C.-W., Senger, S. M., Dickinson, M. E. and dx.doi.org/10.13140/RG.2.2.17778.56005. Udan, R. S. (2018). The effects of reduced hemodynamic loading on morphogenesis of the mouse embryonic heart. Dev. Biol. 442, 127-137. doi:10. Supplementary information 1016/j.ydbio.2018.07.007 Supplementary information available online at Hosseini, H. S., Beebe, D. C. and Taber, L. A. (2014). Mechanical effects of the http://jcs.biologists.org/lookup/doi/10.1242/jcs.230300.supplemental surface ectoderm on optic vesicle morphogenesis in the chick embryo. J. Biomech. 47, 3837-3846. doi:10.1016/j.jbiomech.2014.10.018 References Hosseini, H. S., Garcia, K. E. and Taber, L. A. (2017). A new hypothesis for foregut Aigouy, B., Umetsu, D. and Eaton, S. (2016). Segmentation and quantitative and heart tube formation based on differential growth and actomyosin contraction. analysis of epithelial tissues. Methods Mol. Biol. 1478, 227-239. doi:10.1007/978- Development 144, 2381-2391. doi:10.1242/dev.145193 Hughes, A., Greene, N. D. E., Copp, A. J. and Galea, G. L. (2018). Valproic acid 1-4939-6371-3_13 Alexandre, P., Reugels, A. M., Barker, D., Blanc, E. and Clarke, J. D. W. (2010). disrupts the biomechanics of late spinal neural tube closure in mouse embryos. Mech. Dev. 149, 20-26. doi:10.1016/j.mod.2017.12.001 Neurons derive from the more apical daughter in asymmetric divisions in the Jacobberger, J. W., Frisa, P. S., Sramkoski, R. M., Stefan, T., Shults, K. E. and zebrafish neural tube. Nat. Neurosci. 13, 673-679. doi:10.1038/nn.2547 Soni, D. V. (2008). A new biomarker for mitotic cells. Cytometry A 73A, 5-15. Bush, K. T., Lynch, F. J., DeNittis, A. S., Steinberg, A. B., Lee, H.-Y. and Nagele, doi:10.1002/cyto.a.20501 R. G. (1990). Neural tube formation in the mouse: a morphometric and Kinoshita, N., Sasai, N., Misaki, K. and Yonemura, S. (2008). Apical accumulation computerized three-dimensional reconstruction study of the relationship of Rho in the neural plate is important for neural plate cell shape change and between apical constriction of neuroepithelial cells and the shape of the neural tube formation. Mol. Biol. Cell 19, 2289-2299. doi:10.1091/mbc.e07-12- neuroepithelium. Anat. Embryol. 181, 49-58. doi:10.1007/BF00189727 1286 Cavadino, A., Prieto-Merino, D., Addor, M.-C., Arriola, L., Bianchi, F., Draper, E., Kosodo, Y., Suetsugu, T., Suda, M., Mimori-Kiyosue, Y., Toida, K., Baba, S. A., Garne, E., Greenlees, R., Haeusler, M., Khoshnood, B. et al. (2016). Use of Kimura, A. and Matsuzaki, F. (2011). Regulation of interkinetic nuclear migration hierarchical models to analyze European trends in congenital anomaly by cell cycle-coupled active and passive mechanisms in the developing brain. prevalence. Birth Defects Res. A Clin. Mol. Teratol. 106, 480-488. doi:10.1002/ EMBO J. 30, 1690-1704. doi:10.1038/emboj.2011.81 bdra.23515 Krajcovic, M. M. and Minden, J. S. (2012). Assessing the critical period for Rho Christodoulou, N. and Skourides, P. A. (2015). Cell-autonomous Ca(2+) flashes kinase activity during Drosophila ventral furrow formation. Dev. Dyn. 241, elicit pulsed contractions of an apical actin network to drive apical constriction 1729-1743. doi:10.1002/dvdy.23859 during neural tube closure. Cell Rep. 13, 2189-2202. doi:10.1016/j.celrep.2015. Leitch, C., Osdal, T., Andresen, V., Molland, M., Kristiansen, S., Nguyen, X. N., 11.017 Bruserud, O., Gjertsen, B. T. and McCormack, E. (2016). Hydroxyurea Copp, A., Cogram, P., Fleming, A., Gerrelli, D., Henderson, D., Hynes, A., synergizes with valproic acid in wild-type p53 acute myeloid leukaemia. Kolatsi-Joannou, M., Murdoch, J. and Ybot-Gonzalez, P. (2000). Neurulation Oncotarget 7, 8105-8118. doi:10.18632/oncotarget.6991 and neural tube closure defects. Methods Mol. Biol. 136, 135-160. doi:10.1385/1- Leung, L., Klopper, A. V., Grill, S. W., Harris, W. A. and Norden, C. (2011). Apical 59259-065-9:135 migration of nuclei during G2 is a prerequisite for all nuclear motion in zebrafish Dai, J., Sultan, S., Taylor, S. S. and Higgins, J. M. (2005). The kinase haspin is neuroepithelia. Development 138, 5003-5013. doi:10.1242/dev.071522 required for mitotic histone H3 Thr 3 phosphorylation and normal metaphase Liang, Y.-C., Li, X.-D., Wu, Y.-P., Ke, Z.-B., Liu, Z.-Q., Chen, S.-H., Wei, Y., Zheng, chromosome alignment. Genes Dev. 19, 472-488. doi:10.1101/gad.1267105 Q.-S., Xue, X.-Y. and Xu, N. (2019). LIMK1 depletion enhances fasudil- Das, D., Zalewski, J. K., Mohan, S., Plageman, T. F., VanDemark, A. P. and dependent inhibition of urethral fibroblast proliferation and migration. J. Cell. Hildebrand, J. D. (2014). The interaction between Shroom3 and Rho-kinase is Biochem. 120, 12977-12988. doi:10.1002/jcb.28569 required for neural tube morphogenesis in mice. Biol. Open 3, 850-860. doi:10. Lopez-Escobar, B., Caro-Vega, J. M., Vijayraghavan, D. S., Plageman, T. F., 1242/bio.20147450 Sanchez-Alcazar, J. A., Moreno, R. C., Savery, D., Marquez-Rivas, J., Desmond, M. E. and Jacobson, A. G. (1977). Embryonic brain enlargement Davidson, L. A. and Ybot-Gonzalez, P. (2018). The non-canonical Wnt-PCP requires cerebrospinal fluid pressure. Dev. Biol. 57, 188-198. doi:10.1016/0012- pathway shapes the mouse caudal neural plate. Development 145, dev157487. 1606(77)90364-5 doi:10.1242/dev.157487 Duda, M., Kirkland, N. J., Khalilgharibi, N., Tozluoglu, M., Yuen, A. C., Carpi, N., Macdonald, J. I. and Dick, F. A. (2012). Posttranslational modifications of the Bove, A., Piel, M., Charras, G., Baum, B. and Mao, Y. (2019). Polarization of retinoblastoma tumor suppressor protein as determinants of function. Genes myosin II refines tissue material properties to buffer mechanical stress. Dev. Cell Cancer 3, 619-633. doi:10.1177/1947601912473305 48, 245-260.e7. doi:10.1016/j.devcel.2018.12.020. Martin, A. C., Kaschube, M. and Wieschaus, E. F. (2009). Pulsed contractions of Escuin, S., Vernay, B., Savery, D., Gurniak, C. B., Witke, W., Greene, N. D. E. and an actin-myosin network drive apical constriction. Nature 457, 495-499. doi:10. Copp, A. J. (2015). Rho-kinase-dependent actin turnover and actomyosin 1038/nature07522 disassembly are necessary for mouse spinal neural tube closure. J. Cell Sci. Martino, F., Perestrelo, A. R., Vinarský, V., Pagliari, S. and Forte, G. (2018). 128, 2468-2481. doi:10.1242/jcs.164574 Cellular mechanotransduction: from tension to function. Front. Physiol. 9, 824. Feroze, R., Shawky, J. H., von Dassow, M. and Davidson, L. A. (2015). doi:10.3389/fphys.2018.00824 Mechanics of blastopore closure during amphibian gastrulation. Dev. Biol. 398, Mason, F. M., Tworoger, M. and Martin, A. C. (2013). Apical domain polarization 57-67. doi:10.1016/j.ydbio.2014.11.011 localizes actin-myosin activity to drive ratchet-like apical constriction. Nat. Cell Galea, G. L., Cho, Y. J., Galea, G., Mole, M. A., Rolo, A., Savery, D., Moulding, D., Biol. 15, 926-936. doi:10.1038/ncb2796 Culshaw, L. H., Nikolopoulou, E., Greene, N. D. E. et al. (2017). Biomechanical McGreevy, E. M., Vijayraghavan, D., Davidson, L. A. and Hildebrand, J. D. coupling facilitates spinal neural tube closure in mouse embryos. Proc. Natl. Acad. (2015). Shroom3 functions downstream of planar cell polarity to regulate myosin II Sci. USA 114, E5177-E5186. doi:10.1073/pnas.1619264114 distribution and cellular organization during neural tube closure. Biol. open 4, Galea, G. L., Nychyk, O., Mole, M. A., Moulding, D., Savery, D., Nikolopoulou, E., 186-196. doi:10.1242/bio.20149589 Henderson, D. J., Greene, N. D. E. and Copp, A. J. (2018). Vangl2 disruption McShane, S. G., Mole,̀ M. A., Savery, D., Greene, N. D. E., Tam, P. P. L. and Copp, alters the biomechanics of late spinal neurulation leading to spina bifida in mouse A. J. (2015). Cellular basis of neuroepithelial bending during mouse spinal neural embryos. Dis. Model. Mech. 11, dmm032219. doi:10.1242/dmm.032219 tube closure. Dev. Biol. 404, 113-124. doi:10.1016/j.ydbio.2015.06.003 Guthrie, S., Butcher, M. and Lumsden, A. (1991). Patterns of cell division and Nagasaka, A., Shinoda, T., Kawaue, T., Suzuki, M., Nagayama, K., Matsumoto, interkinetic nuclear migration in the chick embryo hindbrain. J. Neurobiol. 22, T., Ueno, N., Kawaguchi, A. and Miyata, T. (2016). Differences in the mechanical 742-754. doi:10.1002/neu.480220709 properties of the developing cerebral cortical proliferative zone between mice and Harding, M. J. and Nechiporuk, A. V. (2012). Fgfr-Ras-MAPK signaling is required ferrets at both the tissue and single-cell levels. Front. Cell Dev. Biol. 4, 139. doi:10. for apical constriction via apical positioning of Rho-associated kinase during 3389/fcell.2016.00139 mechanosensory organ formation. Development 139, 3130-3135. doi:10.1242/ Nagele, R. G. and Lee, H.-Y. (1979). Ultrastructural changes in cells associated with dev.082271 interkinetic nuclear migration in the developing chick neuroepithelium. J. Exp. Hartmann, S., Ridley, A. J. and Lutz, S. (2015). The function of Rho-associated Zool. 210, 89-106. doi:10.1002/jez.1402100110 kinases ROCK1 and ROCK2 in the pathogenesis of cardiovascular disease. Nair, J. S., Ho, A. L., Tse, A. N., Coward, J., Cheema, H., Ambrosini, G., Keen, N. Front. Pharmacol. 6, 276. doi:10.3389/fphar.2015.00276 and Schwartz, G. K. (2009). Aurora B kinase regulates the postmitotic Hazawa, M., Lin, D. C., Kobayashi, A., Jiang, Y. Y., Xu, L., Dewi, F. R. P., endoreduplication checkpoint via phosphorylation of the retinoblastoma protein

Mohamed, M. S., Hartono, Nakada, M., Meguro-Horike, M. et al. (2018). at serine 780. Mol. Biol. Cell 20, 2218-2228. doi:10.1091/mbc.e08-08-0885 Journal of Cell Science

10 RESEARCH ARTICLE Journal of Cell Science (2019) 132, jcs230300. doi:10.1242/jcs.230300

Nikolopoulou, E., Galea, G. L., Rolo, A., Greene, N. D. E. and Copp, A. J. (2017). Sai, X., Yonemura, S. and Ladher, R. K. (2014). Junctionally restricted RhoA Neural tube closure: cellular, molecular and biomechanical mechanisms. activity is necessary for apical constriction during phase 2 inner ear placode Development 144, 552-566. doi:10.1242/dev.145904 invagination. Dev. Biol. 394, 206-216. doi:10.1016/j.ydbio.2014.08.022 Nishimura, T., Honda, H. and Takeichi, M. (2012). Planar cell polarity links axes of Seller, M. J. and Perkins, K. J. (1983). Effect of hydroxyurea on neural tube defects spatial dynamics in neural-tube closure. Cell 149, 1084-1097. doi:10.1016/j.cell. in the curly-tail mouse. J. Craniofacial Genet. Dev. Biol. 3, 11-17. 2012.04.021 Shinoda, T., Nagasaka, A., Inoue, Y., Higuchi, R., Minami, Y., Kato, K., Suzuki, Oltean, A. and Taber, L. A. (2018). Apoptosis generates mechanical forces that M., Kondo, T., Kawaue, T., Saito, K. et al. (2018). Elasticity-based boosting of close the lens vesicle in the chick embryo. Phys. Biol. 15, 025001. doi:10.1088/ neuroepithelial nucleokinesis via indirect energy transfer from mother to daughter. 1478-3975/aa8d0e PLoS Biol. 16, e2004426. doi:10.1371/journal.pbio.2004426 Oltean, A., Huang, J., Beebe, D. C. and Taber, L. A. (2016). Tissue growth Short, S. P., Kondo, J., Smalley-Freed, W. G., Takeda, H., Dohn, M. R., Powell, constrained by extracellular matrix drives invagination during optic cup A. E., Carnahan, R. H., Washington, M. K., Tripathi, M., Payne, D. M. et al. morphogenesis. Biomech. Model. Mechanobiol. 15, 1405-1421. doi:10.1007/ (2017). p120-Catenin is an obligate haploinsufficient tumor suppressor in s10237-016-0771-8 intestinal neoplasia. J. Clin. Invest. 127, 4462-4476. doi:10.1172/JCI77217 Pearl, E. J., Li, J. and Green, J. B. A. (2017). Cellular systems for epithelial Shum, A. S. W. and Copp, A. J. (1996). Regional differences in morphogenesis of the neuroepithelium suggest multiple mechanisms of spinal neurulation in the invagination. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 372, 20150526. doi:10. mouse. Anat. Embryol. 194, 65-73. doi:10.1007/BF00196316 1098/rstb.2015.0526 Singh, P. N. P., Shea, C. A., Sonker, S. K., Rolfe, R. A., Ray, A., Kumar, S., Gupta, Peeters, M. C. E., Schutte, B., Lenders, M.-H., Hekking, J. W. M., Drukker, J. and P., Murphy, P. and Bandyopadhyay, A. (2018). Precise spatial restriction of BMP Van Straaten, H. W. M. (1998). Role of differential cell proliferation in the tail bud in signaling in developing joints is perturbed upon loss of embryo movement. aberrant mouse neurulation. Developmental dynamics: an official publication of Development 145, dev153460. doi:10.1242/dev.153460 the American Association of Anatomists 211, 382-389. doi:10.1002/(SICI)1097- Smith, J. L. and Schoenwolf, G. C. (1988). Role of cell-cycle in regulating 0177(199804)211:4<382::AID-AJA9>3.0.CO;2-D neuroepithelial cell shape during bending of the chick neural plate. Cell Tissue Philips, F. S., Schwartz, H. S. and Sternberg, S. S. (1968). Hydroxyurea: inhibition Res. 252, 491-500. doi:10.1007/BF00216636 of DNA-synthesis and selective susceptibility of S-phase cells in vivo. Acta Genet. Spear, P. C. and Erickson, C. A. (2012). Apical movement during interkinetic Med. Gemellol. 17, 5-8. doi:10.1017/S1120962300026238 nuclear migration is a two-step process. Dev. Biol. 370, 33-41. doi:10.1016/j. Porazinski, S., Wang, H., Asaoka, Y., Behrndt, M., Miyamoto, T., Morita, H., ydbio.2012.06.031 Hata, S., Sasaki, T., Krens, S. F. G., Osada, Y. et al. (2015). YAP is essential for Vaezi, A., Bauer, C., Vasioukhin, V. and Fuchs, E. (2002). Actin cable dynamics tissue tension to ensure vertebrate 3D body shape. Nature 521, 217-221. doi:10. and Rho/Rock orchestrate a polarized cytoskeletal architecture in the early steps 1038/nature14215 of assembling a stratified epithelium. Dev. Cell 3, 367-381. doi:10.1016/S1534- Ren, J., Tang, C.-Z., Li, X.-D., Niu, Z.-B., Zhang, B.-Y., Zhang, T., Gao, M.-J., Ran, 5807(02)00259-9 X.-Z., Su, Y.-P. and Wang, F.-C. (2018). Identification of G2/M phase transition by Winter, C. G., Wang, B., Ballew, A., Royou, A., Karess, R., Axelrod, J. D. and sequential nuclear and cytoplasmic changes and molecular markers in mice Luo, L. (2001). Drosophila Rho-associated kinase (Drok) links Frizzled-mediated intestinal epithelial cells. Cell Cycle 17, 780-791. doi:10.1080/15384101.2018. planar cell polarity signaling to the actin cytoskeleton. Cell 105, 81-91. doi:10. 1426416 1016/S0092-8674(01)00298-7 Journal of Cell Science

11